Short Circuit Problem

H

hsn6827

Member
Hello everybody,

I have a problem with my Teensy 4.1 board. Today, while i was trying some codes with RC servos, one analog signal wire was touched to servo's GND or +12V mistakenly and caused short circuit. Now, my teensy board does not run. When I energized it via USB, marked component with red rectangle below is becoming hot. What is this component? Any idea? Is it possible to repair it ?
 

Attachments

  • Teensy 4.1 Board.png
    Teensy 4.1 Board.png
    513.1 KB · Views: 230
Last edited:
That part is the voltage regulator. It gets burning hot when another part of the board is shorted (from 3.3V to GND), because the rest of the board is trying to consume nearly limitless current and the regulator tries to work as hard as it can to supply that current.

I'm afraid the news is only bad.

Sometimes a 3.3V-GND short on the board is recoverable when the problem is stray metal, like excess solder. But when the 3.3V-GND short is inside a chip, repair isn't feasible. If you have a voltmeter, you could try measuring the DC voltage from 3.3V to GND. If you see a low number but not close to zero, like 0.5 to 0.9 volts, that's a pretty sure sign the short on your board is a semiconductor short. Metal shorts read zero or only a few millivolts.

Sadly, we've seen this many times with 12 volt projects. Accidentally touching 12 volts to pretty much anywhere, even for only the slightest moment, completely destroys the main processor. It's a sad moment to lose hardware, but at least you can avoid wasting a lot of time on attempting repair and just replace the entire board.
 
PaulStoffregen said:
That part is the voltage regulator. It gets burning hot when another part of the board is shorted (from 3.3V to GND), because the rest of the board is trying to consume nearly limitless current and the regulator tries to work as hard as it can to supply that current.

I'm afraid the news is only bad.

Sometimes a 3.3V-GND short on the board is recoverable when the problem is stray metal, like excess solder. But when the 3.3V-GND short is inside a chip, repair isn't feasible. If you have a voltmeter, you could try measuring the DC voltage from 3.3V to GND. If you see a low number but not close to zero, like 0.5 to 0.9 volts, that's a pretty sure sign the short on your board is a semiconductor short. Metal shorts read zero or only a few millivolts.

Sadly, we've seen this many times with 12 volt projects. Accidentally touching 12 volts to pretty much anywhere, even for only the slightest moment, completely destroys the main processor. It's a sad moment to lose hardware, but at least you can avoid wasting a lot of time on attempting repair and just replace the entire board.
Click to expand...
Thank you very much Paul. When I measure the 3V-GND pins DC voltage, I saw that 0.08 V. It is almostly closer to 0 VDC. So do you have any comment about it?
 
That is pretty close to zero. If you didn't know what went wrong, and especially if the problem had appeared right after soldering pins or doing some other mechanical work where stray metal may have been placed on the circuit board, I would have normally have suggested looking for a metal short.

But you do know what happened. Accidentally touching 12 volts isn't going to make metal magically appear. But it definitely will burn the inside of the main processor. We know this with good certainty because it's happened many times with people building 12 volt projects. The result is always the same, the U1 chip destroyed.

I know it's bad news, but hopefully not as painful as wasting a lot more time trying fix something that has virtually no hope of repair. It's time to just replace the damaged hardware.

Probably not very comforting, but plenty of other folks working with 12V have had similar mishaps. I've even done a couple times. Usually after something like this happens it's a good idea to think about connectors or cable (zip) ties or other ways to keep higher voltage wires secured while working on the hardware. Just 1 brief errant moment is all it takes to completely destroy the main chip.
 
Tips for a bulletproof setup:
  1. Use a poly fuse at least for low current applications, or a miniature circuit breaker for more demanding amps.
  2. When dealing with higher voltages than the MCU, the safest thing to do is use optical isolation. Both digital and analog options exist, at all sorts of speeds.
  3. Use a crowbar circuit to immediately cut off the voltage if it goes too high. This is what I have been doing on my projects/products that deal with different voltages. it's very simple to construct. It works by shorting supply to ground, and that's where the fuse/breaker comes into play.
  4. Use a diode to prevent the usual stupid mistake that I make, to prevent hooking up power backwards. I usually set it up so it will trip the fuse like the crowbar does, or at the input if there's a regulator following the input.
 
PaulStoffregen said:
That is pretty close to zero. If you didn't know what went wrong, and especially if the problem had appeared right after soldering pins or doing some other mechanical work where stray metal may have been placed on the circuit board, I would have normally have suggested looking for a metal short.

But you do know what happened. Accidentally touching 12 volts isn't going to make metal magically appear. But it definitely will burn the inside of the main processor. We know this with good certainty because it's happened many times with people building 12 volt projects. The result is always the same, the U1 chip destroyed.

I know it's bad news, but hopefully not as painful as wasting a lot more time trying fix something that has virtually no hope of repair. It's time to just replace the damaged hardware.

Probably not very comforting, but plenty of other folks working with 12V have had similar mishaps. I've even done a couple times. Usually after something like this happens it's a good idea to think about connectors or cable (zip) ties or other ways to keep higher voltage wires secured while working on the hardware. Just 1 brief errant moment is all it takes to completely destroy the main chip.
Click to expand...
Thank you for your suggestion and advice. As you mentioned or estimated in your post, yes I didn't use connector to connect my servos to power supply. It was possible risk for short circuit. Thus it happened. Ok. No need to spend time for recover. It is better to new one even it is expensive :(
 
xxxajk said:
Tips for a bulletproof setup:
  1. Use a poly fuse at least for low current applications, or a miniature circuit breaker for more demanding amps.
  2. When dealing with higher voltages than the MCU, the safest thing to do is use optical isolation. Both digital and analog options exist, at all sorts of speeds.
  3. Use a crowbar circuit to immediately cut off the voltage if it goes too high. This is what I have been doing on my projects/products that deal with different voltages. it's very simple to construct. It works by shorting supply to ground, and that's where the fuse/breaker comes into play.
  4. Use a diode to prevent the usual stupid mistake that I make, to prevent hooking up power backwards. I usually set it up so it will trip the fuse like the crowbar does, or at the input if there's a regulator following the input.
Click to expand...
Thank you very much for your very beneficial advices @xxxajk . They are very important tips. I have added some of them like using diodes to prevent reverse polarity troubles. But I will add fuse and crowbar circuit to circuit to increase protection.
Additionally to this list, do not any wiring or hardware changing without cutting energy from the circuit. :)
 
hsn6827 said:
Thank you very much for your very beneficial advices @xxxajk . They are very important tips. I have added some of them like using diodes to prevent reverse polarity troubles. But I will add fuse and crowbar circuit to circuit to increase protection.
Additionally to this list, do not any wiring or hardware changing without cutting energy from the circuit. :)
Click to expand...
You are quite welcome. If you need a schematic for the crowbar circuit, let me know.
 
hsn6827 said:
Yes, if you have an example, could you share it with me please?
Click to expand...
Sure, here's one.
NOTES:
SCR trip voltage is 400mV, you may need to adjust R1 to a lower value to decrease sensitivity.

Code: 
volts    D1        R1    trips at
3.3v     1N5333     60K    3.49V
5.0v     1N5228     60K    6.09V
Screenshot_20250523_104817.png
 
xxxajk said:
Sure, here's one.
NOTES:
SCR trip voltage is 400mV, you may need to adjust R1 to a lower value to decrease sensitivity.

Code: 
volts    D1        R1    trips at
3.3v     1N5333     60K    3.49V
5.0v     1N5228     60K    6.09V
View attachment 37606
Click to expand...
Thank you very much @xxxajk . But if you have full circuit, could you share it please? I could not understand that how i can place this diagram to my circiut.
 
Funny thing just happened to one of my teensy 4.1's, was running fine, and then regulator hot, and not working anymore. Nothing plugged into it, just running PSRAM and sdcard tests, and it just "gave up" and died apparently.
SDcard is fine, have to check PSRAM, but I suspect it is OK too.
 
xxxajk said:
Funny thing just happened to one of my teensy 4.1's, was running fine, and then regulator hot, and not working anymore. Nothing plugged into it, just running PSRAM and sdcard tests, and it just "gave up" and died apparently.
SDcard is fine, have to check PSRAM, but I suspect it is OK too.
Click to expand...
I wondered the reason. Why it can be mulfunction without any plugged thing. I think Teensy boards are very sensitive.
 
3.3V high frequency devices are more sensitive than the 5V 16MHz ATMEGA328 used on standard Arduinos.
I/Os going to the "outside world", i.e. connected through connectors and long cables to peripherals like encoders, servos, sensors,..., should be protected: 100ohms series resistors, RC filters, diodes, .....
Teensy 3.3V power supply should not be used for these devices, unless it is protected agains EMI comming back to it. Here also, use ferrites, coils, capacitors, diodes,....
 
In this case after positioning trouble for RC servo, marked component is becoming hot. It is a microcontroller about USB communication. Now PC can not recognize it. How can I repair it if it is possible? Can I replace microcontroller without any diagnostic? @PaulStoffregen could you check it please? Do you have any idea?
 

Attachments

  • Teensy_4.1.png
    Teensy_4.1.png
    302.1 KB · Views: 98
hsn6827 said:
In this case after positioning trouble for RC servo, marked component is becoming hot. It is a microcontroller about USB communication. Now PC can not recognize it. How can I repair it if it is possible? Can I replace microcontroller without any diagnostic? @PaulStoffregen could you check it please? Do you have any idea?
Click to expand...
@xxxajk , do you have any suggestion about this issue? How can i protect and isolate my teensy board for the PWM output pins and analog input pins? For pwm, i can use optocoupler module. Is it logical?
 
hsn6827 said:
@xxxajk , do you have any suggestion about this issue? How can i protect and isolate my teensy board for the PWM output pins and analog input pins? For pwm, i can use optocoupler module. Is it logical?
Click to expand...
Yeah, I usually do that, just make sure you get one that is able to respond quick enough to the signalling.
100 ohms to the LED to limit current is pretty much ideal for most of them.
You also need a pull-up resistor on the output, as it will be open collector, 1K to 2.2K depending on what you have, and the current requirement.
You can also make the output non-inverting by flipping the pin that you drive on the LED if it is important to the servo. For some this is required. Drive it on the negative side instead of the positive side. That way the LED turns on during the low phase, turning on the NPN transistor inside the optoisolator and you get a low on that side too.
 
xxxajk said:
Yeah, I usually do that, just make sure you get one that is able to respond quick enough to the signalling.
100 ohms to the LED to limit current is pretty much ideal for most of them.
You also need a pull-up resistor on the output, as it will be open collector, 1K to 2.2K depending on what you have, and the current requirement.
You can also make the output non-inverting by flipping the pin that you drive on the LED if it is important to the servo. For some this is required. Drive it on the negative side instead of the positive side. That way the LED turns on during the low phase, turning on the NPN transistor inside the optoisolator and you get a low on that side too.
Click to expand...
Thank you @xxxajk . I designed 2 diffrent circuits below. PWM signal protecitons are same for both of them with optocoupler. But for the analog signal pins protection, 1st one was made according to your protection circuit advice. Components are not totally same with your circuit since I could not find same in here. So I have to changed them with alternatives which are existing in here. I am not sure if they run normally or not. :)
2nd one is my own design with TVS diode. Could you check them please? Do you have any suggestion or idea? I dont want to waste my microcontroller anymore :D
 

Attachments

  • 1st_Circuit.png
    1st_Circuit.png
    72.2 KB · Views: 77
  • 2nd_Circuit.png
    2nd_Circuit.png
    70.3 KB · Views: 73
You can do analog by using an opto isolator that has PIN diodes and an opamp.
Basically it allows you to do analog "level shifting" of sorts.
Also the PTC is on the wrong side of the circuit...
 
Last edited:
xxxajk said:
You can do analog by using an opto isolator that has PIN diodes and an opamp.
Basically it allows you to do analog "level shifting" of sorts.
Also the PTC is on the wrong side of the circuit...
Click to expand...
Thank you @xxxajk . I tried to use 817 optocoupler component on my PWM signal line. Not worked. May be frequecy is not enough. I will try high frequency types.

Feedback signal voltage comes from Servo to analog pin. So when the voltage exceed the protection limit, short circuit will be. To protect line, I should cut voltage by PTC. So it should be on line voltage source side. Am I right? If not, could you explain its logic please?
 
hsn6827 said:
Thank you @xxxajk . I tried to use 817 optocoupler component on my PWM signal line. Not worked. May be frequecy is not enough. I will try high frequency types.

Feedback signal voltage comes from Servo to analog pin. So when the voltage exceed the protection limit, short circuit will be. To protect line, I should cut voltage by PTC. So it should be on line voltage source side. Am I right? If not, could you explain its logic please?
Click to expand...
The idea for the PCT is for voltage supply to the MCU, not as an input.
For analog signal input from something not in the 3.3v range, you want to optically isolate using an optoisolator with analog compensation PIN diode.
VOA300-DEFG-X017T and the IL300 are interchangeable, I've used both with the following circuit. Select the one that best meets your timing requirements.
 

Attachments

  • opto.png
    opto.png
    39.7 KB · Views: 59
Other notes:
You shouldn't need anything faster than 1MHz ability for PWM.

The PTC circuit is meant to protect the MCU power supply, it does not reset unless powered off.
The idea is to short an over voltage source to ground, and the PTC takes the abuse of the short circuit that is protecting your MCU...
For example, plugging in 12V into a 5v circuit because you grabbed the wrong adapter.
 
You must log in or register to reply here.
Back
Top